A major question in diabetes research is the relationship of improved glucose control to the reduction of long-term complications of the disease. There is presently no way to obtain good glucose control in diabetics over the sustained periods required to answer this question. We propose a development program to build a programmable insulin delivery system designed for external use on ambulatory animals and humans. It could be the forerunner of a implantable unit for chronic health care. The system will be small (fist size), light-weight (about 300 grams), low power (2-year battery life-time), and reliable (our goal - one failure or less per 100 unit-years operation). It will use a proven, high-performance, rotary-solenoid-driven peristaltic pump. It will use standard CMOS-technology-integrated circuits to provide a wide variability of basal and postprandial delivery rates. This approach represents a low-risk development program to establish the basic technology for a smaller implantable unit using large-scale integrated circuitry when appropriate in the future. The system will deliver from microliters to milliliters per hour of insulin to give maximum flexibility to the biomedical experiments which can be performed. The system's performance will be assessed in both biomedical and laboratory experiments. The biomedical experiments will use the somatostatin-induced diabetic dog. In these experiments, flow/concentration/ absorption relationships will be determined using intravenous, intraperitoneal, and subcutaneous insulin delivery modes.